"According to their latest blurb, Sheepshaver supports up to MacOS 8.1 _only_. Critically, even when it is running on a Mac it is useless for multimedia apps and games: it will _not_ run applications which access the hardware directly.

I tried the Sheepshaver demo on my Starmax running BeOS 3.2 and it hung constantly. I don't think it is at all ready for prime time.

The IBM design is interesting, as is the recently released Motorola "Linux" machine. But until Be supports these (remotely possible), or Apple allows Be to run on G3 and later machines (verging on impossible), or Apple supports these machines directly (another universe), ...... grrr I am almost as put off by Jobs little monolith as I am with Gates...:-) "

Another reader commented on the IBM reference design motherboards:

"Two issues...
1) The CHRP Motherboard (Longtrail) design by IBM is extensively documented (I can make the developement documentation available on request)

IBM states in "Long Trail: The PowerPC Platformfor MacOS and Beyond"

"Several operating systems are or will be supported ... being used in the development of CHRP MacOS 1.0. Long Trail is also a regression platform used by Apple Computer for the development of Tempo, or MacOS 8.0."

In this case, a MacOS ROM is required. IBM provides all the pieces (schematics, bill of materials, specifications) required to make the ROM DIMM as a normal burned ROM or as a FLASH ROM DIMM (which would be most valuable for the purpose of making MAC OS boot on IBM's new Linux motherboard design.)

2) However, none of this may be necessary. The current version of MacOS supports loading the MacOS ROM from hard disk via Open Firmware. I have had mixed success getting a IBM Power Series 830 to boot MacOS (More on this as it develops).

Anonymously Yours... "

News.com has a new article on the cost of Windows 2000, citing an estimated 'migration' cost of up to $3,100. I've ran a beta copy (as part of their corp. preview program) and was disappointed in the reliability and performance (both CPUs were shown as 50% usage at idle, crashes, horribly slow booting). The hopes of NT's better stability/performance and Win9x's multimedia support quickly vanished within hours of the Win2000 beta3 install. I won't go into the details of needing motherboard flash ROM and driver updates. Release Candidate 1 may address some of these issues but I cringe at the thought of another install. What's most depressing is apparently it still has 16-bit code.

With OS 9 to have (I hear) core support for Altivec and the release of the new G4 systems I think Apple has a real window of opportunity to gain more market share. Even developers here seem frustrated at the maze of future windows versions. I'm trying now to get a local company to look at the G4 systems instead of PCs with an AMD Athlon CPU. I'm hoping my G4/500 will arrive before they make a final decision. They're looking hard at a Kryotech active cooled 800MHz Athlon system that has very impressive performance. One of my concerns on the Athlon is the possibility of undiscovered flaws in the new CPU and/or motherboard chipset needed for the 'slot A' Athlon CPU. The Athlon as noted here previously has some very advanced features and usually far outperforms the Pentium III.

G4 Performance Editorial: Several readers sent a note on an article commenting on the new Apple G4 systems (some saw the link at Macfixit). First let me say that I think the new G4/AGP Macs are by far the best Macs ever made and the best values ever. They have improvements in literally every area (better memory bus, an AGP slot, ATA/66, better USB). I have not had the pleasure to test a G4 system (yet) but I want to make some comments about realistic expectations based on what I've seen in testing a G4 vs G3 CPU and comparing those results to a B&W G3. Many buyers are expecting 'supercomputer' performance increases over their existing G3s even with current applications based on all the marketing claims.

I have to say comments that a G4 CPU in an B&W G3 is like 'strapping a rocket booster to a skate board' is a bit of an exaggeration in my opinion. Soon there will be results of Yikes (basically BW G3 design) vs Sawtooth in real world apps and we'll see how the 'skateboard' really stacks up. Regardless, I'd personally not buy the Yikes if I could afford the improved Sawtooth design (Maxbus, AGP, etc.).

Based on non-Altivec apps tests results and Motorola's own docs it is clear that the G4 core is similar to the G3 with a better FPU (like the 604e). The G4's claim to fame is the Altivec execution engine; and with most of today's apps that engine is idle. This is verbatum from the Motorola docs on the G4 CPU talking about non-Altivec performance:

"MPC7400's [G4] core is essentially the same as the MPC750's [G3], except that whereas the MPC750 has a 6-entry completion queue and has slower performance on some floating-point double-precision operations, MPC7400 has an 8-entry completion queue and a full double-precision FPU."

For instance I saw no real difference in Bryce or Infini-D performance between a G4/400 and a G3/400 (50MHz or 100MHz bus speeds made literally no difference, so don't let anyone say the memory bandwidth was a bottleneck on this test). But if you listen to some of the claims being made on the G4, you'd expect it to blow the G3 away at any task. As of this date in time that simply is not true with most of the programs you are using now. (3D Games and Photoshop 5.5 with the G4 plugins will be exceptions.)

Since non-Altivec optimized apps are what 99% of end users will be running when their G4 arrives, I think some of marketing claims are putting unrealistic expectations into many consumers heads. Specs and memory bandwidth are not linear indicators of real world performance. For instance - the B&W G3 has twice the bus speed and more than twice the memory bandwidth of my Genesis system - but is the B&W really any faster at most apps I use than the Genesis with the same speed CPU? Not usually, and in most cases they perform within 1-2% of each other (documented here in many reviews and articles). Bandwidth specs, bus speeds, max. transfer rates all look impressive on paper but many common applications today aren't really limited by memory bandwidth (at least with a 1MB backside cache buffering main memory). I'm very eager to see what improvements the AGP/Maxbus Sawtooth provides, especially with apps that use Altivec instructions (I'd expect up to 4x improvements in this case).

As seen on the PC, AGP vs PCI is another common case of the specifications being more impressive than typical observed performance benefit. There are cases where AGP's higher bandwidth, direct to memory bus (not shared with the PCI bus) can show dramatic gains but these are usually the exception rather than the rule. Again most Mac software is written to not tax even the 132MB/sec max of the standard 33MHz PCI bus. And in most cases your video card is already a bottleneck (it must have the fill rate to take advantage of higher bandwidth).

PC AGP vs PCI tests in many 3D games show the difference is often not visible and certainly far less than the interface specs. See this AGP vs PCI article at the well-respected Tom's Hardware site for more info. Would I prefer AGP? You bet! But I'm not foolish enough to think that I'm going to see dramatic benefits from it with today's Mac software. First I'd need a video card that has a fill rate that is not already saturated on the current B&W G3s (the Rage128 already is as seen in my Quake 2 hi-res game tests). The problem (and many marketers know this) is that many consumers see '528MB/sec AGP' and think it's got to be 4 times as fast as '132 MB/sec' std. PCI. Don't get me wrong - I've been praying for AGP to come to the Mac (regular readers will remember my January 1999 complaint that the B&W G3 didn't have AGP). AGP is the current industry standard for graphics cards and is one of the major reasons I bought a Sawtooth G4 system. Again an investment in the future.

As more software and the OS becomes Altivec optimized the benefits of owning a G4 will increase, but I'm talking about most common applications people are using today. For these apps I'll bet a blind test of comparabily equipped G4 vs G3 systems would have users hard pressed to 'see' any real difference. There are many other benefits to the new systems that make them attractive but new buyers shouldn't expect their old apps to run dramatically faster than a similarly configured G3 system. Buy a new G4 system for the many other improvements and new features in addition to the G4 processor. Consider it an investment in the future.

I'm very excited about the G4/Altivec and new AGP/G4 systems, and only post this long-winded commentary as I feel there are many people out there with unrealistic expectations about basic (non-Altivec) G4 CPU performance. If you can afford it, by all means get the new AGP G4 system. With Altivec optimizations targeted for the OS, Quicktime, etc. you'll see increasing benefits in the future. The Adobe Altivec plugins that ship with the new systems will pay off immediately for Photoshop 5.5 owners. I agree 100% that the new systems are far superior designs and only want to caution the 'typical' end user that already owns a fast G3 to have realistic expectations on performance with most applications they're currently using.

Also remember that performance isn't the only criteria - consider all factors (features, ease of use, support, reliability, your software needs) before making any purchasing decision. Often new systems (at least the first revision) bring their own set of compatibility issues that can affect your productivity. (I admit I'm a performance addict and rarely make decisions based on genuine need and never follow my own advice of waiting for 'rev 2' designs, but running this site is my excuse for spending far too much time and money on computers. :-)

After first posting this I realized I'm forgotten to say that the AGP/G4 systems addressed literally every facet of my wish list (better bus design, AGP slot, ATA/66, improved USB, provisions for multiple processors) and still used affordable components (PC100 RAM, IDE drives). I am very pleased with the design of the new Sawtooth Macs and can't wait to get mine.

BTW: According to the specs I have unlike stated elsewhere, the Pentium III is not a 7x hotter (temperature) CPU than the G4. The specs I have for the G4/400MHz says 8 Watts max. The current Pentium III 450MHz specs cite 25.3 Watts, but they include the 512K L2 cache and bus terminating resistors in that figure.

More TrueX SCSI Feedback: I've received another reader report on the Kenwood TrueX SCSI CDROM drive:

"Just installed the new Kenwood 52x-True-x CD-Rom (SCSI connection) for my Umax S900 computer and it works incredibly with my Intech CD-Rom speedtools driver (I can't get it to work with just the Apple CD-Dvd driver). I've tested it with audio files, SoundJam, and Toast 3.5.7 and it works flawlessly. Its speed makes you think you're working with a hardrive. I love your site.
Respectfully Yours,
Patrick Adiarte"

He later reported the some similar quirks with the CD Audio player application as I saw in testing the IDE TrueX drive. Other players seem fine, including the control strip audio cd player.

Falcon 4 Pilots Wanted: Noel sent a note he is looking for Mac Falcon Pilots to fly with and to get a tournament off the ground.

"To All Mac Pilots.
Maestro and I are starting a Mac Pilots competition ladder/Tournament. Email me if anyone is interested, Maestro is hard a work designing the web site. The current address is http://sites.netscape.net/jbentony/ If anyone has any ideals please let me know. Email me if you have any questions or concerns. iway@epix.net "

Jay Craft sent some interesting comments on using a reference design motherboard with the MacOS:

"Mike, just a thought...
I was talking with someone on IRC about the IBM/Motorola reference powerpc motherboard designs, and the fact that since they would not contain apple ROMs, they couldn't run MacOS. However, Sheepshaver (http://www.sheepshaver.com) is a MacOS "environment" for Beos and soon LinuxPPC that can boot MacOS (with a snapshot of a real mac ROM image) and run at supposedly full speed with great compatibility. Note that a BeBox, which can run Sheepshaver, probably has nothing in common with a Mac other than that it uses a powerPC as its cpu... It would seem that in this way someone could likely build a cheap very fast box that runs MacOS (in some form)...

Which leads me to the next thought: I wonder if you could boot LinuxPPC on a G3 (and eventually G4) "without ROM" upgraded, imac or Powerbook G3, and then get your MacOS work done with something like Sheepshaver... Obviously not something for the average user but just a thought- if you can boot MacOS 8.6 on a non-mac PowerPC machine, surely someone could figure out how to do it on an actual Mac without the hardware ROM.
Jay Craft"

Interesting. I'm sure it could be done - but I wonder if Apple allow it. Apple could crush most small companies with legal action if they saw this as a threat. Ignoring that, once you had the ROM image (not legal to use perhaps unless you owned a Mac - the issues of this sort of thing are beyond my legal knowledge) - it would be possible I'd think to use a bootstrap loader to load the rom at startup. But the Mac ROM is designed to interface to Mac motherboards (memory map), so I think there would need to be some other layer of software for compatibility with the reference designs perhaps (just a guess - this might not be an issue). What would be nice is a utililty like System Commander on the PC (or a modified Linux Lilo perhaps) - where at boot a loader gives you a menu to select which OS to load/run (a selection of say Linux, BeOS, MacOS 8/9, or OS X would be great wouldn't it).

I'd get one of the ref designs in a minute personally - even if I had to run BeOS/Sheepsaver. I wonder if Sheepsaver will be updated to run OS 9 or X?

USB/Clean OS Install Tip: I received this tip on clean OS installs and USB devices:

"The following is not just a Mac OS 9 issue -it's universal.
After a clean install of any system, devices that are plugged into USB host adapters will no longer function. The 1.2 USB libraries must be reinstalled to correct the problem. "

A reader sent a link to the latest Apple Tech Exchange forum reply on the B&W G3 ROM check. This link was valid as of the today but the post may be deleted in the future. BTW, I hear the ROM version in the 'Yikes' (low end G4) is Firmware v1.2f2, ROM v1.7 [corrected].

I'm testing the Formac Proformance III now in several systems for a review. As I've commented before, its a very nice card with excellent software controls, very fast 2D and decent game performance. I'm having to retest with the recent ROM and control panel updates for the card. OpenGL drivers were expected this month and will round out the driver support for the card. The 3D Glasses option is a nice addition and really adds impressive effects in Rave Quake and other games. I've have a full review and comparison to cards like the Rage128 Orion and 3Dfx Voodoo3 in the review.

Allan Landwer wrote that the XLR8 MAChCarrier G3/450 card and software cured his previous problem with RAM Disk copy corruption {SpeedDoubler 8.1.2 copy verify failed) in his PowerTowerPro 250 he saw with his original G3 upgrade card (an early PowerForce G3/275). I asked if he had tested the PF G3 card with Powerlogix's latest software and he said yes but that only XLR8's 1.4 control panel fixed the problem. XLR8 revised the CarrierZIF card shortly after introduction to address specific problems with the PowerTower Pro 250 motherboard. For more info on Allan's system and upgrades, see his web site.

I've added a speed settings chart/tips to my CarrierZIF Setup Guide and divided the article into linked pages as the guide is necessarily image heavy.

Gene Steinberg's latest Arizona Central column covers his recent interview with Microsoft's Dick Craddock about their upcoming Macintosh software

theiMac.com has a new article on iBook Battery Conservation. My friend Peter Bethke is due to get his iBook soon and I'll try to get him to write a review for the site (how about it Peter? ;-).

ZDNet has a story today on Internet Explorer 5 for the Mac delayed, citing work on the page rendering engine as reasons why. BTW: While at Macworld I spoke to the MS reps there regarding improving Mac Office 98 performance, asking them to look into the reasons why Mac Excel is so much slower than the PC version (my guess was the Mac code may be some sort of VB interpreter). For examples of this see the performance tests I did last year between a PB G3/250 and Gateway Solo notebook.

If you don't check every day's daily news (updated several times a day normally) you might miss something valuable. Check the summary list below for highlights of the last few day's news and the archives page for older new page links. For a blast from the past, read what was on the news page here exactly a year ago.

Wednesday's News Summary

• More OWC Specials
• USB/SCSI Adapter Feedback
• G4 Yikes Systems Arriving
• G4 vs G3 CPU Tech Details
• MP3 Encoding from CD vs HD
• Update on the B&W ROM G4 CPU Check
• ZIF CPU Gasket Note
• More on USB/ADB Adapters
• RAM Prices Rise Again
• ZDNET reviews the eOne
• ORB performance comparison
• Mac Game Framerates database updated
• Rate Your G3 Upgrade database updated
• Rate Your Mac Games database

Tuesday's News Summary

• Adjustable ZIFs & Non-Standard Bus Speeds
• ATI Extends Trade-Up Rebate
• Fast G3s/Birmy RIP Tip
• Rage128 Pro in some G4s?
• More G4 CPU Tech Info
• G4 Forum Discussions
• Another 3Dfx Interview discusses future products
• PowerBook Multi-LCD Spoof
• More on USB/ADB Adapters
• G4 Systems topic area added to Freq. Asked Questions
• BTV 3.0 Video Utility released
• G4 vs G3 performance tests updated
• Mac Game FPS database updated
• Rate Your Mac Games database updated
• Rate Your G3 Upgrade database updated

Monday's News Summary

• Voodoo3 AGP in Sawtooth (3dfx comments existing ROM should work)
• Errors converting Audio track to AIFF in Soundjam MP v1.1
• PhotoShop G4 Plugins - only available w/G4 systems?
• B&W G3 ROM Flashing - attempts to flash 1.0x version fail
• My initial tests on OWC 350/400 ZIFs in B&W G3
• G3 Logic Board Upgrades for very old Macs
• ZIF Heatsink Clip Reminder (orientation important for cooling)
• Problems downloading Techworks Power3D (Voodoo 1) drivers
• Apple's Stock Hits Record High
• Game News page updated
• Mac Game Framerates database updated
• Rate Your G3 Upgrade database updated
• Rate Your Mac Games database

For a guide to finding answers to questions - see my Site Guide page.

If you own a G3/300 or faster Mac, you might want to get our PPC Checker utility that will report if your CPU is copper or aluminum based.

Considering a G3 Upgrade? Check the CPU Upgrade Owner's Survey results and search for owner reports by Mac model and/or card brand. The database is being updated daily for new reports. Also make sure you ask your dealer about any issues with your particular mac model as I have gotten several reports of issues with certain G3 upgrades in some Mac models (Mach 5, PowerBase, 7300). I urge any owners that have had problems to report this via the Survey Form. Always give the most weight to recent owner reports (dated in the listing).

-Freq Asked Questions